Get Fulloch

Fulloch runs entirely on your own machine — wake word, speech-to-text, language model and voice, all on-device. Getting it going is quick: grab the code and run one script.

You’ll need an NVIDIA GPU with 16GB of VRAM (an RTX 5060 Ti is the reference card) and either Linux or Windows (with Docker Desktop + Python 3.10+).

1. Get the code

#

Clone the repo:

git clone https://github.com/liampetti/fulloch.git
cd fulloch

Or download it as a ZIP and unzip it.

2. Launch it

#

Linux:

./launch.sh

Windows: (install Python 3.10+ and Docker Desktop first)

pip install -r requirements.txt
launch.bat

That’s it — the launcher creates your config, downloads the models, sets up the audio, and starts everything. When it’s ready, say “Hey Atticus” and start talking.

Full setup details, Home Assistant integration and configuration options are all in the repo README.